Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A robotic camera system for inventory monitoring, comprising: a movable base; multiple cameras supported by the movable base, the multiple cameras being directable toward inventory; a processing module connected to the multiple cameras and able to construct inventory related information, the processing module configured to: stitch together consecutive images from the multiple cameras into a panorama, segment the panorama to define a plurality of bounding boxes, classify the plurality of bounding boxes into different categories, associate bounding boxes among different categories based on geographic position of the plurality of bounding boxes, and determine the inventory-related information based on the association of the bounding boxes; a communication module connected to the processing module to transfer the inventory related information to remote locations; and a navigation module connected with the movable base, the communication module, and the multiple cameras to direct the position and orientation of the robotic camera system.
A robotic camera system monitors inventory in storage environments. The system addresses challenges in accurately tracking and categorizing inventory items, particularly in large or dynamic storage spaces where manual monitoring is inefficient. The system includes a movable base that supports multiple cameras, which can be directed toward inventory items. A processing module connected to the cameras constructs inventory-related information by stitching consecutive images from the cameras into a panoramic view. The panoramic image is then segmented to define multiple bounding boxes, each representing a detected inventory item. These bounding boxes are classified into different categories based on visual features, and the system associates bounding boxes from different categories based on their geographic positions within the panorama. This association helps determine inventory-related information, such as item counts, locations, and relationships between items. A communication module transfers this inventory data to remote locations for further analysis or record-keeping. A navigation module controls the movement and orientation of the robotic system, ensuring optimal camera positioning for accurate inventory monitoring. The system automates inventory tracking, reducing human error and improving efficiency in warehouse or storage management.
2. The camera system of claim 1 , wherein the processing module being further configured to create an updateable realogram from the associated bounding boxes.
A camera system captures images or video of a scene and processes the data to identify and track objects within the field of view. The system includes a processing module that generates bounding boxes around detected objects in the captured images or video frames. These bounding boxes define the spatial boundaries of each object, allowing the system to track their positions and movements over time. The processing module is further configured to create an updateable realogram, which is a dynamic representation of the scene that incorporates the bounding boxes. The realogram can be updated in real-time as new objects are detected or existing objects move, providing an evolving spatial map of the scene. This allows for applications such as object tracking, scene analysis, and augmented reality, where real-time spatial awareness is critical. The system may also include additional features, such as object classification, motion prediction, or integration with other sensors, to enhance its functionality. The updateable realogram ensures that the system maintains an accurate and current representation of the scene, enabling precise tracking and interaction with objects in real-time.
3. The camera system of claim 1 , wherein the movable base further comprises an autonomous robot.
The invention relates to a camera system designed for dynamic and autonomous surveillance or imaging applications. The system addresses the challenge of capturing images or video from multiple vantage points without manual repositioning, which is inefficient and labor-intensive. The camera system includes a movable base that supports a camera module, allowing the camera to be repositioned as needed. The movable base is equipped with an autonomous robot, enabling it to navigate and reposition itself independently based on predefined instructions or environmental inputs. This robot may use sensors, navigation algorithms, or remote control to move the camera to desired locations, ensuring optimal coverage or tracking of subjects. The camera module itself may include features such as adjustable lenses, zoom capabilities, or multiple sensors to enhance image quality and flexibility. The autonomous robot may also incorporate obstacle avoidance, path planning, or real-time adjustments to ensure smooth and efficient movement. This system is particularly useful in applications like security monitoring, environmental observation, or industrial inspections where continuous, adaptable imaging is required.
4. The camera system of claim 1 , wherein the multiple cameras are linearly mounted on a camera support.
A camera system is designed to capture high-resolution images or video by using multiple cameras arranged in a specific configuration. The system addresses the challenge of achieving high-resolution imaging without the cost and complexity of a single large sensor. By using multiple smaller cameras, the system can capture detailed images while maintaining a compact and scalable design. The cameras are linearly mounted on a camera support, which ensures precise alignment and synchronization of the individual camera modules. This linear arrangement allows for seamless stitching of the captured images or video frames, creating a high-resolution output. The camera support provides structural stability and may include mechanisms for adjusting the position or orientation of the cameras to optimize image capture. The system may also include processing components to synchronize the cameras, correct for distortions, and combine the captured data into a single high-resolution output. This approach is particularly useful in applications requiring detailed imaging, such as surveillance, medical imaging, or industrial inspection.
5. The camera system of claim 1 , wherein at least one of the multiple cameras is fixedly mounted on a camera support.
A camera system is designed to capture images or video from multiple perspectives, addressing the need for enhanced situational awareness, surveillance, or environmental monitoring. The system includes multiple cameras that can be positioned to cover different angles or areas, improving coverage and reducing blind spots. At least one of these cameras is fixedly mounted on a camera support, which provides stability and precise positioning. The support ensures the camera remains in a designated orientation, preventing unwanted movement and maintaining consistent imaging quality. This fixed mounting is particularly useful in applications requiring long-term monitoring, such as security systems, industrial inspections, or autonomous vehicle navigation. The system may also include additional cameras that are adjustable or movable, allowing for dynamic adjustments to capture changing scenes or targets. The fixed camera ensures a reliable reference point, while other cameras can adapt to varying conditions. The overall design enhances flexibility and reliability in capturing comprehensive visual data.
6. The camera system of claim 1 , wherein at least one of the multiple cameras is a zoom camera.
A camera system includes multiple cameras configured to capture images of a scene from different perspectives. The system is designed to enhance image capture by providing multiple viewpoints, which can be used for applications such as 3D imaging, depth sensing, or panoramic photography. The inclusion of at least one zoom camera allows for variable magnification, enabling the system to capture detailed images of distant objects while maintaining a wide field of view with other cameras. The zoom camera can be adjusted to focus on specific regions of interest within the scene, while the remaining cameras provide additional context or complementary data. This configuration improves flexibility in capturing high-quality images under varying conditions, such as low light or dynamic environments. The system may also incorporate image processing techniques to combine data from multiple cameras, enhancing overall image quality and depth perception. The zoom functionality allows for dynamic adaptation to different imaging requirements without the need for physical movement of the entire system, making it suitable for applications in surveillance, robotics, or autonomous vehicles.
7. The camera system of claim 1 , wherein at least one of the multiple cameras is directed downward and at least one of the multiple cameras is directed upward.
A camera system is designed to capture images or video from multiple directions, particularly for applications requiring comprehensive environmental monitoring. The system includes multiple cameras arranged to provide overlapping or complementary fields of view. At least one camera is oriented downward to capture images of the ground or lower surfaces, while at least one other camera is oriented upward to capture images of the sky or upper surfaces. This configuration allows the system to monitor both upward and downward directions simultaneously, which is useful for applications such as surveillance, environmental monitoring, or autonomous navigation. The downward-facing camera may be used to track ground-based objects or terrain, while the upward-facing camera may detect aerial objects or weather conditions. The system may also include additional cameras positioned at different angles to provide a more complete view of the surrounding environment. The arrangement of cameras ensures that no critical direction is left unmonitored, improving situational awareness and safety. The system may be integrated into vehicles, drones, or fixed installations to provide real-time data for decision-making or analysis.
8. The camera system of claim 1 , further comprising a plurality of lights directable toward inventory.
A camera system for inventory management includes a camera configured to capture images of inventory items and a processor that analyzes the images to identify and track inventory. The system addresses challenges in accurately monitoring stock levels, item locations, and inventory conditions in warehouses or retail environments. To enhance visibility and imaging quality, the system incorporates a plurality of lights that can be directed toward the inventory. These lights improve image clarity by illuminating items, reducing shadows, and ensuring consistent lighting conditions for accurate identification and tracking. The processor may adjust lighting parameters, such as intensity or direction, based on environmental factors or inventory characteristics to optimize image capture. The system may also include additional features like motion detection, automated alerts for low stock, and integration with inventory management software. By combining advanced imaging with adaptive lighting, the system provides real-time, reliable inventory data to improve operational efficiency and reduce errors in stock management.
9. The camera system of claim 1 , wherein the multiple cameras include at least one of a wide field camera, a fixed camera, and a movable camera having a resolution sufficient to read shelf attached inventory information.
A camera system is designed for monitoring retail environments, particularly for tracking inventory and shelf conditions. The system includes multiple cameras with different capabilities to capture comprehensive visual data. These cameras may include a wide field camera for broad area coverage, a fixed camera for stable, high-resolution imaging of specific sections, and a movable camera that can adjust its position or orientation to capture detailed images of shelf-attached inventory information, such as barcodes or labels. The movable camera is equipped with sufficient resolution to accurately read and interpret this inventory data, enabling real-time or periodic updates to inventory management systems. The combination of these camera types allows the system to monitor large retail spaces efficiently while ensuring precise tracking of individual items. This setup helps automate inventory management, reduce manual counting efforts, and improve accuracy in stock monitoring. The system is particularly useful in retail stores, warehouses, or automated fulfillment centers where maintaining up-to-date inventory records is critical.
10. The camera system of claim 1 , wherein the multiple cameras include an infrared camera.
A camera system is designed to capture images or video in various lighting conditions, including low-light or nighttime environments. The system includes multiple cameras, at least one of which is an infrared camera. Infrared cameras detect heat signatures and emit infrared light to illuminate scenes, enabling visibility in darkness or obscured conditions. The system may also include other types of cameras, such as visible light cameras, to provide complementary imaging data. The infrared camera enhances the system's ability to capture clear images in low-light scenarios, improving situational awareness and security monitoring. The system may be used in surveillance, automotive, or industrial applications where reliable imaging in challenging lighting conditions is essential. The integration of infrared imaging with other camera types allows for versatile use in environments where traditional visible light cameras would fail to provide adequate visibility.
11. An inventory monitoring method, comprising the steps of: allowing an autonomous robot to move along a shelf lined aisle holding inventory, with the autonomous robot acting as a movable base for multiple cameras; directing the multiple cameras toward inventory and shelf labels on the shelf lined aisle; stitching together consecutive images from the multiple cameras into a panorama; segmenting the panorama to define a plurality of bounding boxes; classifying the plurality of bounding boxes into different categories comprising shelf labels, inventory labels, and inventory; associating bounding boxes among different categories based on geographic position of the plurality of bounding boxes; determining inventory-related information based on the associations of the bounding boxes; and transferring the inventory-related information to remote locations using a communication module.
This invention relates to automated inventory monitoring in warehouse or retail environments. The problem addressed is the need for efficient, accurate, and real-time tracking of inventory levels and locations, which is traditionally labor-intensive and prone to human error. The solution involves an autonomous robot equipped with multiple cameras that moves along aisles lined with shelves holding inventory. The cameras capture images of inventory items and shelf labels, which are then stitched together to form a panoramic view of the aisle. The panoramic image is processed to segment and classify different elements, including shelf labels, inventory labels, and the inventory items themselves, using bounding boxes to delineate each detected object. The system associates these bounding boxes based on their geographic positions, enabling the determination of inventory-related information such as item locations, quantities, and potential discrepancies. This data is then transmitted to remote locations via a communication module, allowing for centralized inventory management and analysis. The method improves inventory accuracy, reduces manual labor, and enhances operational efficiency in large-scale storage or retail settings.
12. An inventory monitoring method, comprising the steps of: allowing an autonomous robot to move along a shelf lined aisle holding inventory, with the autonomous robot acting as a movable base for multiple cameras; maintaining a substantially constant distance from the shelf lined aisle holding inventory while moving in a forward direction; directing the multiple cameras toward inventory on the shelf lined aisle; generating a panorama from images captured by the multiple cameras; segmenting the panorama to define a plurality of bounding boxes; classifying the plurality of bounding boxes into categories comprising shelf labels, inventory labels, and inventory; associating bounding boxes among different categories based on the geographic position of the bounding boxes; and determining inventory-related information based on the association of bounding boxes.
This invention relates to automated inventory monitoring in warehouse or retail environments. The problem addressed is the need for efficient, accurate, and scalable methods to track inventory levels, identify misplaced items, and detect missing or incorrect labels without manual intervention. Traditional inventory management systems often rely on fixed cameras or manual scanning, which are time-consuming and prone to errors. The method involves an autonomous robot equipped with multiple cameras that moves along aisles lined with shelves holding inventory. The robot maintains a consistent distance from the shelves while advancing forward, ensuring optimal image capture. The cameras are directed toward the inventory, capturing images that are stitched together to form a panoramic view of the aisle. This panorama is then segmented into multiple bounding boxes, each representing distinct objects or regions of interest. The bounding boxes are classified into three categories: shelf labels, inventory labels, and inventory items. The system then associates bounding boxes across these categories based on their geographic positions, allowing it to correlate inventory items with their corresponding labels and shelf locations. Finally, the system determines inventory-related information, such as stock levels, label accuracy, and item placement, based on these associations. This automated approach improves inventory accuracy, reduces labor costs, and enhances operational efficiency in large-scale storage or retail facilities.
13. The camera system of claim 12 , wherein the multiple cameras are linearly mounted on a camera support.
A camera system is designed to capture high-resolution images or video by using multiple cameras arranged in a specific configuration. The system addresses the challenge of achieving high-resolution imaging without the cost and complexity of a single high-resolution camera. By using multiple lower-resolution cameras, the system can capture a wider field of view or higher detail while maintaining cost efficiency. The cameras are mounted on a support structure to ensure precise alignment and synchronization. In this particular configuration, the multiple cameras are linearly mounted on a camera support, meaning they are arranged in a straight line. This linear arrangement allows for seamless stitching of images or video from adjacent cameras, reducing gaps and improving overall image quality. The support structure ensures stability and proper positioning of the cameras, which is critical for accurate image capture and alignment. The system may also include processing components to combine the outputs from the multiple cameras into a single high-resolution image or video stream. This approach is particularly useful in applications such as surveillance, aerial imaging, and high-resolution photography where a wide field of view or high detail is required.
14. The camera system of claim 12 , wherein at least one of the multiple cameras is fixedly mounted on a camera support.
A camera system is designed to capture images or video from multiple perspectives, addressing the need for comprehensive visual data in applications such as surveillance, autonomous navigation, or environmental monitoring. The system includes multiple cameras configured to capture overlapping or adjacent fields of view, ensuring seamless coverage of a target area. At least one of these cameras is fixedly mounted on a camera support, which provides structural stability and precise positioning. The support may be adjustable or integrated into a larger assembly, allowing for optimal alignment and orientation of the camera. The fixed mounting ensures consistent image quality and reduces vibration or movement that could distort captured data. Additional cameras in the system may be mounted similarly or in alternative configurations, depending on the specific application requirements. The system may also include processing components to merge or analyze the captured data, enhancing situational awareness or enabling advanced functionalities such as object tracking or 3D reconstruction. The fixed mounting of at least one camera ensures reliable operation in dynamic environments, making the system suitable for industrial, security, or scientific use.
15. An inventory monitoring method, comprising the steps of: allowing an autonomous robot to move along a shelf lined aisle holding inventory, with the autonomous robot acting as a movable base for multiple cameras; directing the multiple cameras toward inventory on a shelf of the shelf lined aisle; detecting distance from at least one of the multiple cameras to the shelf using a distance ranging system; determining inventory related-information from camera images, including: generating a panorama of the shelf by stitching consecutive images captured by the multiple cameras; segmenting the panorama to define a plurality of bounding boxes; classifying the plurality of bounding boxes into categories comprising shelf labels, inventory labels, and inventory; associating bounding boxes among different categories based on the geographic location of the plurality of bounding boxes; and determining the inventory-related information based on the associated of the bounding boxes; and inferring real time inventory status of the shelf from the inventory-related information, including at least one of: integrity of a product location or on-shelf stock status of a product.
This invention relates to automated inventory monitoring in retail or warehouse environments, addressing the challenge of accurately tracking stock levels and product placement without manual intervention. The system employs an autonomous robot equipped with multiple cameras as a movable base, navigating along aisles lined with shelves. The cameras capture images of inventory on the shelves, while a distance ranging system measures the robot's proximity to the shelves. The captured images are processed to generate a panoramic view of the shelf by stitching consecutive frames. This panorama is then segmented into bounding boxes, which are classified into categories such as shelf labels, inventory labels, and inventory items. The system associates these bounding boxes based on their geographic locations, enabling the extraction of inventory-related information. By analyzing the relationships between the classified boxes, the system infers real-time inventory status, including product location integrity and on-shelf stock levels. This automated approach improves efficiency and accuracy in inventory management, reducing the need for manual checks and minimizing stockouts or misplacements.
16. The inventory monitoring method of claim 15 , wherein generating a panorama of the shelf comprises generating an updateable realogram of the shelf; and wherein segmenting the panorama to define a plurality of bounding boxes comprises segmenting the panorama using one of: machine learning or deep learning.
This invention relates to inventory monitoring systems that use computer vision to track and analyze product placement on retail shelves. The problem addressed is the inefficiency of manual inventory checks and the need for automated, real-time monitoring of shelf stock levels and product arrangement. The method involves capturing images of a retail shelf using one or more cameras, then generating a panorama of the shelf from these images. The panorama is an updateable realogram, meaning it is a digital representation of the shelf layout that can be modified as products are added, removed, or rearranged. The system segments this panorama into multiple bounding boxes, each representing a distinct product or product group, using either machine learning or deep learning techniques. These segmentation methods allow the system to accurately identify and track individual items on the shelf. The segmented panorama is then analyzed to detect discrepancies between the current shelf state and a predefined ideal arrangement, such as missing products, incorrect placements, or misaligned items. The system can generate alerts or reports to notify store personnel of these discrepancies, enabling timely restocking and proper product placement. The use of machine learning or deep learning ensures high accuracy in product recognition and segmentation, reducing errors in inventory monitoring. This automated approach improves efficiency, reduces labor costs, and enhances customer satisfaction by maintaining well-stocked and properly organized shelves.
17. The inventory monitoring method of claim 15 , further comprising using the distance ranging system to capture a depth map.
This invention relates to inventory monitoring systems that use distance ranging technology to track and manage items in a storage environment. The system addresses challenges in accurately detecting and locating items within a space, particularly in dynamic or cluttered environments where traditional methods like barcodes or RFID may be unreliable. The method involves deploying a distance ranging system, such as a LiDAR or depth-sensing camera, to scan the environment and generate spatial data representing the positions and dimensions of objects. This data is processed to identify and classify items based on their physical characteristics, such as size and shape, and to track their movement over time. The system may also incorporate machine learning algorithms to improve recognition accuracy and adapt to changes in the inventory. Additionally, the method includes capturing a depth map, which provides a three-dimensional representation of the scanned environment, enhancing the precision of item localization and reducing errors caused by occlusions or overlapping objects. The system can be integrated with inventory management software to update stock levels automatically, trigger reorder alerts, or optimize storage layouts. This approach improves efficiency in warehouses, retail stores, and other inventory-intensive settings by automating tracking and reducing manual counting efforts.
18. The inventory monitoring method of claim 15 , further comprising using the distance ranging system to capture a depth map including both shelves and products positioned on shelves.
The invention relates to inventory monitoring systems that use depth-sensing technology to track products in retail or warehouse environments. The problem addressed is the need for accurate, real-time inventory tracking without manual scanning or disruptive sensors. The system employs a depth-ranging system, such as a LiDAR or structured light sensor, to capture three-dimensional spatial data of shelves and products. This data is processed to generate a depth map, which represents the positions and dimensions of products on shelves. The depth map is analyzed to identify product locations, quantities, and potential stocking errors. The system may also compare the captured data against a reference model to detect discrepancies, such as missing or misplaced items. By continuously monitoring inventory in this way, the system enables automated stock management, reducing the need for manual audits and improving supply chain efficiency. The depth-ranging system can be integrated into existing retail infrastructure, such as ceiling-mounted sensors or mobile scanning devices, to provide scalable and non-intrusive monitoring. The method ensures accurate inventory tracking by leveraging depth information to distinguish between products and shelf structures, improving reliability over traditional 2D imaging systems.
19. The inventory monitoring method of claim 15 , further comprising building a product library.
This invention relates to inventory monitoring systems, specifically addressing the challenge of efficiently tracking and managing product data in large-scale retail or warehouse environments. The method involves dynamically updating an inventory database by capturing product information from various sources, such as barcodes, RFID tags, or manual input, and correlating this data with existing records. A key feature is the ability to detect discrepancies between scanned items and expected inventory, triggering alerts for potential stock issues like theft, misplacement, or supply chain errors. The method further includes building a product library, which serves as a centralized repository of standardized product identifiers, descriptions, and attributes. This library ensures consistency across different inventory systems and facilitates faster data retrieval. The library may be populated using manufacturer databases, supplier catalogs, or crowdsourced data, and it can be updated in real-time as new products are introduced or existing ones are modified. The system also supports automated categorization of products based on predefined criteria, such as product type, brand, or expiration date, to streamline inventory management. By integrating real-time monitoring with a structured product library, the invention improves inventory accuracy, reduces manual data entry errors, and enhances decision-making for restocking and logistics. The solution is particularly useful in high-volume retail environments where rapid and precise inventory tracking is critical.
20. The inventory monitoring method of claim 15 , further comprising building a product library using manual input to identify potential inventory designated in the multiple bounding boxes in the image taken with at least one camera.
This invention relates to automated inventory monitoring systems that use image processing to track and identify products in a storage or retail environment. The core challenge addressed is the accurate and efficient detection and classification of inventory items from images captured by cameras, particularly in dynamic or cluttered settings where manual tracking is impractical. The method involves capturing images of an inventory area using one or more cameras. The system then processes these images to detect and isolate individual products by generating multiple bounding boxes around potential inventory items. To enhance accuracy, the method includes building a product library through manual input, where users or operators manually identify and label the products within the bounding boxes. This manual input helps train the system to recognize and classify inventory items more reliably in subsequent automated scans. The product library serves as a reference database, enabling the system to cross-check detected items against known product profiles. This step improves the system's ability to distinguish between similar-looking products or handle variations in lighting, orientation, or partial occlusions. The manual input process may involve tagging, categorizing, or specifying attributes for each product, ensuring the system can accurately track inventory levels, detect discrepancies, or trigger restocking alerts. By combining automated image processing with manual verification, the method aims to balance efficiency and accuracy in inventory management, reducing human labor while maintaining high reliability in product identification. This approach is particularly useful in warehouses, retail stores, or automated fulfillment centers where real-time inventory tracking
21. The inventory monitoring method of claim 15 , wherein the bounding boxes can surround gaps between potential inventory.
This invention relates to inventory monitoring systems, specifically methods for detecting and tracking inventory items using bounding boxes in visual data. The core problem addressed is accurately identifying and monitoring inventory, including cases where items are partially obscured or separated by gaps. The method involves analyzing visual data, such as images or video, to detect potential inventory items. Bounding boxes are generated around these items to define their spatial boundaries. A key feature is the ability to surround gaps between potential inventory items with a single bounding box, allowing for more accurate tracking of closely spaced or partially obscured items. This is particularly useful in environments where items are stored in close proximity or where partial visibility occurs. The system may use computer vision techniques, such as object detection algorithms, to identify inventory items and generate bounding boxes. The method ensures that even when items are not fully visible or are separated by small gaps, the bounding boxes still encompass the relevant areas, improving inventory tracking accuracy. This approach helps in scenarios like warehouse management, retail stock monitoring, or automated inventory systems where precise detection is critical. The invention enhances existing inventory monitoring by providing a more robust way to handle complex arrangements of items, reducing errors in counting and tracking. The use of bounding boxes that can include gaps ensures that no inventory is missed, even in cluttered or partially obscured environments.
22. The inventory monitoring method of claim 15 , further comprising building the product library without use of an initial planogram.
The invention relates to inventory monitoring systems, specifically methods for tracking and managing product inventory in retail environments without relying on pre-existing planograms. Traditional inventory monitoring systems often depend on planograms—predefined layouts of products on shelves—to track stock levels. However, these systems can be inflexible, requiring manual updates when product arrangements change. The invention addresses this limitation by providing a method that dynamically builds a product library without an initial planogram, allowing for real-time inventory tracking and adjustments. The method involves capturing images or sensor data of retail shelves, processing the data to identify and classify products, and constructing a product library based on the detected items. Machine learning or computer vision techniques may be used to recognize products, their locations, and quantities. The system can then monitor inventory levels over time, detect stockouts, and alert staff when restocking is needed. By eliminating the need for a predefined planogram, the method adapts to changes in product placement, reducing manual intervention and improving accuracy. The system may also integrate with point-of-sale data to further refine inventory tracking. This approach enhances efficiency in retail operations, particularly in environments where product layouts frequently change.
Unknown
September 8, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.